Learners' and Teachers' Trust, Emotions, and Agency in AI‐Mediated Language Education

ABSTRACT The rapid integration of artificial intelligence (AI) into English as a Foreign Language (EFL) classrooms has reshaped feedback, assessment, and interactional practices. However, limited research has examined how learners' trust in AI systems relates to their emotional experiences and agency in AI‐mediated instruction. This study investigated the structural relationships among EFL translation learners' trust in AI‐assisted language feedback and assessment, foreign language enjoyment, foreign language classroom anxiety, learner autonomy, and perceived control. A cross‐sectional survey design was employed with 732 university EFL translation learners who had prior experience using AI‐supported language learning tools. Data were analysed through structural equation modelling (SEM) using AMOS. The findings revealed that trust in AI‐assisted feedback significantly and positively predicted foreign language enjoyment, learner autonomy, and perceived control, while negatively predicting foreign language classroom anxiety. Furthermore, foreign language enjoyment positively predicted learner autonomy and perceived control, whereas foreign language classroom anxiety exerted significant negative predictive effects on both outcomes. Trust also showed indirect effects on learner autonomy and perceived control through emotional pathways. The findings indicate that learners' trust in AI‐assisted feedback is associated with both their emotional experiences and agency‐related outcomes in AI‐mediated EFL learning. These findings provide empirical evidence for understanding the relationships among trust in AI, achievement emotions, learner autonomy, and perceived control in AI‐supported language learning environments.